Xref: utzoo comp.arch:7798 comp.misc:4653
Path: utzoo!attcan!uunet!lll-winken!ames!amdahl!nsc!glennw
From: glennw@nsc.nsc.com (Glenn Weinberg)
Newsgroups: comp.arch,comp.misc
Subject: Re: "big endian" and "little endian" - first usage for computer
Summary: NS32K is almost completely consistent
Message-ID: <8999@nsc.nsc.com>
Date: 9 Jan 89 22:47:41 GMT
References: <20264@ames.arc.nasa.gov> <2015@scolex>
Reply-To: glennw@nsc.nsc.com.UUCP (Glenn Weinberg)
Organization: National Semiconductor, Sunnyvale
Lines: 25

In article <2015@scolex> seanf@scolex.UUCP (Sean Fagan) writes:
>In article <20264@ames.arc.nasa.gov> lamaster@ames.arc.nasa.gov.UUCP (Hugh LaMaster) writes:
>>I could be wrong, but I think a fully consistent little-endian machine
>>(e.g. nsc 32xxx) does not have this disadvantage.
>
>You're wrong.  On a NSC 32k, addresses are in the wrong order (actually, I
>think it might just be displacements), because the upper 1 or 2 bits
>determine the size of the address (and means that you can't use a
>displacement of 2gigs unsigned, or 1 gig signed.  everybody sigh in unison
>8-)).  Also, I'd bet that the FP format is backwards (wrt big vs. little
>endian).

He's less wrong than you are :-)  The 32K is completely consistently
little-endian (including floating-point), except for displacements,
which are as you described: the upper two bits of the displacement
determine whether it is one, two or four bytes long.  Since displacements
are part of the instruction stream rather than the data, all data
representations in the 32K are consistently little-endian.  Unless you
write self-modifying code, the only time the reverse order of the
displacements is annoying is when you're writing an assembler or
disassembler.

-- 
Glenn Weinberg					Email: glennw@nsc.nsc.com
National Semiconductor Corporation		Phone: (408) 721-8102
(My opinions are strictly my own, but you can borrow them if you want.)